Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Initialization and training of matrix product state probabilistic models

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Born machines trained by gradient descent converge to a causal model that omits the boundary interaction; natural-gradient descent and a sketch-based warm start both escape this trap.

desk verdict Useful diagnosis of a real Born-machine training failure, but the 'trap' claim needs longer runs and convergence evidence; the optimization remedy is honestly reframed as 1-site DMRG. read the letter →

arxiv 2505.06419 v1 pith:U2LA7ZVM submitted 2025-05-09 math.NA cs.NAquant-ph

Xun Tang , Yuehaw Khoo , Lexing Ying This is my paper · ORCID
classification math.NAcs.NAquant-ph MSC 65K10
keywords BornmachinematrixproductstatecausalitytrapnaturalgradientdescentquantumtomographyTTNS-Sketchtensornetworkgenerativemodelsnegativelog-likelihood
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that randomly initialized matrix product state (MPS) models trained by gradient descent can fail at a basic probabilistic modeling task: learning a periodic spin distribution. In the Born machine setting, the trained model converges not to the true distribution $p^\star\propto\exp(-\beta\sum_{(i,j)\in\mathrm{cycle}(n)}x_ix_j)$ but to a causal model $p_{\mathrm{causal}}\propto\exp(-\sum_{(i,j)\in\mathrm{path}(n)}x_ix_j)$ that omits the boundary edge between sites 1 and $n$, and this failure persists at internal rank 20. The paper identifies this as a causality trap, gives a diagnostic for detecting it from the negative log-likelihood gap and the mutual information of the boundary pair, and proposes two cures: a natural gradient descent method that approximates projected gradient flow on the tensor manifold, and a TTNS-Sketch warm initialization. If correct, this means standard gradient training of MPS-based generative models is unreliable for periodic or long-range-correlated targets, and that the failure is curable by the right optimization geometry or a good starting point.

What carries the argument

The load-bearing object is the causality-trap diagnostic: a trained Born machine is classified as trapped when its NLL gap over the optimum equals the mutual information $I(X_1;X_n)$ of the target distribution and the mutual information of the same pair under the model is near zero, conditions equivalent to $D_{w|z}(p_\theta)\approx 0$. The paper's cure is a natural gradient step whose curvature term measures change in the tensor $q_\theta$ rather than in the parameters, which makes the update an approximate projection of $\nabla_q F$ onto the tangent space of one tensor component; Algorithm 1 sweeps sites forward and backward and is equivalent to gradient descent in a mixed canonical form, i.e. a 1-site DMRG update. The warm start uses the TTNS-Sketch algorithm to obtain a direct MPS density estimate $\hat p_{TS}$, takes its positive square root, interpolates it by TT-cross, and fits the resulting MPS to the Born machine ansatz.

What would settle it

Train a randomly initialized Born machine with gradient descent on the periodic Ising distribution with $n=16$, $\beta=1$, and $|T|=2^{15}$; if the trained model's total variation distance to $p^\star$ is smaller than its distance to $p_{\mathrm{causal}}$, or if the NLL gap is not approximately $I(X_1;X_n)\approx 0.33$, the causality-trap claim fails. A finite-$n$ version with all $2^n$ configurations enumerated would settle the same question exactly without sampling noise.

Watch

Extended reading notes

Core claim

The central claim is that the training dynamics of a Born machine under gradient descent on a periodic Ising target converge to the wrong model class: the learned density is close to $p_{\mathrm{causal}}$ in total variation, not to $p^\star$. The paper characterizes the trap through the decomposition $D_{\mathrm{KL}}(p^\star\|p)=D_z(p)+D_{w|z}(p)$ with $z=(x_1,x_n)$ and $w=(x_2,\ldots,x_{n-1})$; trained models satisfy $D_{w|z}(p_\theta)\approx 0$ while $D_z(p_\theta)\approx I_{X\sim p^\star}(X_1;X_n)$, so the NLL gap equals the boundary mutual information. The same local-minimum phenomenon appears in MPS quantum state tomography on a periodic transverse-field Ising ground state. Natural gradient descent, defined as the site-wise minimization of $\langle\nabla_\theta L,\delta\theta\rangle + \tfrac{1}{2\eta}\|q_{\theta_t+\delta\theta}-q_{\theta_t}\|_F^2$, is shown to be a discretization of a projected gradient flow and reaches the global minimum in both settings, with line search needing only a few iterations; warm-starting gradient descent with a TTNS-Sketch density estimate avoids the trap even when that estimate uses only 128 samples.

Load-bearing premise

The whole diagnosis rests on the assumption that two easy-to-measure aggregate statistics, the NLL gap and the mutual information of the boundary pair, are a faithful stand-in for the harder structural condition that the model has learned the wrong causal graph; that equivalence is proven only in the infinite-sample limit and is approximate, yet it is used to classify the finite-sample runs.

Editorial extensions

If this is right

  • For periodic spin targets, a randomly initialized Born machine trained by gradient descent should not be assumed to capture long-range boundary correlations; the NLL gap and the boundary-pair mutual information are enough to tell whether it has fallen into the trap.
  • Natural gradient descent, implemented as single-tensor-component updates with a function-space curvature term, reaches the global minimum in the Born machine and tomography experiments, and with line search does so within a few iterations, so the local-minimum problem is an artifact of the optimization geometry, not of the MPS ansatz.
  • Over-parameterization alone does not cure the trap: increasing $r_{\max}$ from 2 to 20 leaves the trained model at the causal NLL level.
  • A warm initialization obtained from a cheap density estimator, even one fed only 128 samples, is enough for gradient descent to escape the trap, so initialization quality matters more than sample-hungry pretraining.
  • Because the natural gradient derivation is not tied to gauge choices, the same update rule extends to tensor network families where gauge transformations are unavailable, such as positive MPS and locally purified states.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the causality-trap diagnosis is as sharp as the paper suggests, the same NLL-gap-equals-mutual-information test could be used as a stopping or rejection criterion in any MPS training run, flagging models whose boundary correlations are unreliable before expensive total-variation checks.
  • The trap is formulated for the cycle graph, but the mechanism—a chain-like approximation that drops the closing edge—is plausible for any high-dimensional distribution with one strong long-range or non-local interaction; testing on other graphical models with a single non-local bond would show whether the phenomenon is generic.
  • The warm-start result hints that a cheap direct-MPS density estimate plus local fine-tuning could replace expensive random-restart training across tensor-network generative models, but extending it to general quantum state tomography needs a density-estimation subroutine for observables other than computational-basis samples, which the paper leaves open.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper studies training of matrix product state (MPS) probabilistic models, both Born machines (BM) and general MPS quantum state tomography (QST). It reports that gradient descent (GD) from random initialization gets stuck in suboptimal solutions, specifically a 'causality trap' in which the trained BM approximates a path-graph Ising model that omits the boundary edge of the true cycle graph. The paper proposes two fixes: a natural gradient descent (NGD) method derived from a projected gradient flow on the MPS manifold, and a warm-start initialization based on the TTNS-Sketch density estimator. Numerical experiments on periodic Ising and Heisenberg models show that GD fails while NGD, especially with line search, converges in very few iterations, and that warm-started GD avoids the causality trap even with few sketching samples.

Significance. If the reported failure mode is real, the paper identifies a practically relevant pathology of MPS-based generative models and offers two actionable remedies, one of which (NGD) is derived cleanly from a gradient-flow perspective and connects to the TDVP literature. The KL decomposition in Appendix A and the NGD derivation in Appendix C are mathematically sound, and the warm-start strategy leverages a published sample-complexity guarantee for TTNS-Sketch. However, the central negative claim depends on interpreting flat NLL curves as convergence to local minima, and the empirical support consists entirely of single-run experiments without convergence diagnostics or error bars. With appropriate strengthening, the manuscript would be a useful contribution to the numerical analysis of tensor-network training.

major comments (4)
  1. [Section II B, Figures 3-5] The claim that GD converges to a suboptimal local minimum (or to the causality trap) is not substantiated because no convergence diagnostics are reported. The flat NLL curves after 1000 iterations are interpreted as evidence of a local minimum, yet the paper provides no gradient norms, no Hessian information, no convergence criterion, and no longer runs. This matters directly because Appendix B shows that GD on a multilinear MPS ansatz can have an effective learning rate multiplied by (prod_{k≠i} c_k)^2, which can be exponentially small; a plateau over 1000 iterations is fully consistent with an exponentially slow transient rather than a critical point. Please report gradient-norm trajectories, continue until a first-order optimality condition is met, or otherwise demonstrate that the iterates are at a local minimum. Without this, the 'causality trap' as a convergence phenomenon is not established.
  2. [Appendix A, Eqs. (A1)-(A2)] The proposed diagnostic for the causality trap is logically incomplete. The conditions D_w|z(pθ)≈0 and D_z(pθ)≈I_{X~p*}(X1;Xn) are claimed to be verifiable from the NLL gap and the mutual information I_{pθ}(X1;Xn) being near zero. However, from DKL(p*||pθ) = D_z(pθ)+D_w|z(pθ) and small I_{pθ}(X1;Xn), one cannot conclude D_z(pθ)≈I_{p*}(X1;Xn): pθ_{x1,xn} could be independent yet have marginals that differ substantially from p*_{x1,xn}, making D_z(pθ) arbitrarily different from the mutual information of p*. The paper uses this diagnostic to classify finite-sample training runs as trapped, so the inference needs to be justified, e.g., by directly estimating the conditional KL terms or by showing that the trained marginal pθ_{x1,xn} is close to that of pcausal.
  3. [Section IV C, Figures 6-8] The headline quantitative claim that 'NGD with line search can converge to the global minimum in only a few iterations' is supported only by single-run experiments with no error bars, no repeated initializations, and no stated stopping tolerance. Given that the paper's central contrast is between GD failure and NGD success, it is important to show that the results are robust across random initializations and over a reasonable range of the learning rate η and line-search settings. Please provide multiple runs (e.g., mean and standard deviation over seeds) or at least report the specific hyperparameters and tolerances used.
  4. [Appendix C, Eq. (C1)] The proof of Proposition 1 uses Π_i = M(M^T M)^{-1} M^T, which requires M to have full column rank. This is not guaranteed for arbitrary tensor components, and the paper itself notes that trained MPS models exhibit rank degeneracy (Section II B). At a rank-deficient M, the NGD step is undefined by this expression. Please either state a regularity assumption that is maintained by the algorithm (e.g., via gauge fixing or truncation) or replace the inverse with a pseudo-inverse and discuss the rank-deficient case.
minor comments (5)
  1. [Section IV B] The text says 'Theorem 1 is directly related' and refers to 'Theorem 1' twice, but no Theorem 1 is stated; the statement appears to be Proposition 1. Please renumber or cross-reference consistently.
  2. [Appendix E] Appendix E refers to 'Theorem 2' and 'Theorem 3' without having introduced them in the main text; the main text only states Proposition 2. Please align the numbering between the main text and the appendix.
  3. [Section IV A] The statement that 'running Algorithm 1 has only a time complexity of O(n)' is ambiguous: it is unclear whether this is the cost per site update, per full sweep, or per iteration, and the dependence on the bond dimension r is omitted. Please clarify the complexity statement.
  4. [General] The experimental sections do not report the learning rate η used for GD, NGD, or the 2-site DMRG method, nor do they discuss sensitivity to this hyperparameter. Reporting these values would improve reproducibility.
  5. [Figure 3] In Figure 3(b), the y-axis is on a logarithmic scale with values down to 10^{-14}; the caption does not mention the log scale. Please note this in the caption for readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivations are self-contained and the empirical claims are not constructed from their own outputs.

full rationale

I walked the paper's derivation chain. The causality-trap diagnosis in Section II and Appendix A is not a prediction generated from fitted inputs: Appendix A derives, via the standard KL chain rule, the identity D_KL(p*||p_causal) = D_z(p_causal) ≈ I_{X~p*}(X1,Xn), and then uses the independently evaluated NLL gap and mutual-information diagnostics as approximate evidence that the trained p_theta lies near p_causal. Those diagnostics are not parameters fitted to the target and then renamed as predictions; they are computed from the trained model and compared with independent reference values. The natural-gradient method in Section IV is derived from a projected gradient flow in state space, with Proposition 1 proved by direct computation in Appendix C, and the toy example in Appendix B explicitly shows the mechanism by which GD can stall while NGD does not; no target quantity is used as an input. Section V's warm-start proposal cites the TTNS-Sketch sample-complexity guarantee from prior published work [11]. Although [11] shares authors with this paper, it is an externally published result with its own proof, and the numerical demonstration that the warm initialization avoids the causality trap is shown directly, even at |S|=128 samples, so the empirical claim does not reduce to the citation. The only self-citation is [11], and it is not load-bearing for the central failure-mode claim. I found no equation in which the claimed output is defined in terms of itself, no fitted input called a prediction, and no uniqueness theorem imported from the authors' prior work to force a choice. Therefore the paper is self-contained against its own benchmarks and deserves a circularity score of 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim does not introduce new physical entities or fitted constants. The experiments rely on standard MPS math, the representability of the Ising model by low-rank MPS, the TTNS-Sketch theorem from prior work, and the assumption that the discrete NGD updates track the continuous gradient flow. The learning rate and bond dimensions are chosen by hand but not fully reported, which matters for reproducibility.

free parameters (4)
  • learning rate eta = not reported
    NGD and GD updates use eta as the step size; the reported convergence behavior depends on its value, which is not stated in the paper.
  • maximal internal rank r_max = 2, 4, 20 (BM); 6, 10, 20 (tomography)
    Model complexity hyperparameter; the over-parameterization claims rely on these chosen values.
  • sample sizes |T| and |S| = |T| = 2^15; |S| = 2^7 to 2^13
    Experimental setup; the success of warm start with |S| = 128 is a data-dependent observation.
  • beta and system size n = beta = 1, n = 16 or 20
    Parameters chosen for the periodic Ising and TFIM models; the failure mode may depend on these values.
assumptions (4)
  • domain assumption The true distribution p* in Eq (4) is exactly representable by a BM ansatz with r_max <= 4.
    Used to claim the failure is not due to limited expressivity; the model class contains the target.
  • domain assumption TTNS-Sketch algorithm (from [11]) provides a density estimator with sample complexity O~(n^{2+epsilon}/epsilon^2) for BM-representable distributions.
    Proposition 2 relies on the theorem in [11]; the proof is not reproduced in this paper.
  • domain assumption The continuous-time gradient flow in Eq (8) is well-posed and Algorithm 1 is a valid forward Euler discretization.
    The NGD method's convergence properties are analyzed in this continuous limit, not for the discrete updates.
  • standard math Standard mathematical results: KL divergence chain rule, MPS multilinearity, projection formula.
    Used in Appendix A and C; no new mathematical framework is introduced.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Initialization and training of matrix product state probabilistic models." pith.science (2026). https://pith.science/paper/U2LA7ZVM

@misc{pith2026250506419,
  author       = {Pith},
  title        = {Pith review of: Initialization and training of matrix product state probabilistic models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U2LA7ZVM}},
  note         = {Machine review of arXiv:2505.06419}
}
read the original abstract

Modeling probability distributions via the wave function of a quantum state is central to quantum-inspired generative modeling and quantum state tomography (QST). We investigate a common failure mode in training randomly initialized matrix product states (MPS) using gradient descent. The results show that the trained MPS models do not accurately predict the strong interactions between boundary sites in periodic spin chain models. In the case of the Born machine algorithm, we further identify a causality trap, where the trained MPS models resemble causal models that ignore the non-local correlations in the true distribution. We propose two complementary strategies to overcome the training failure -- one through optimization and one through initialization. First, we develop a natural gradient descent (NGD) method, which approximately simulates the gradient flow on tensor manifolds and significantly enhances training efficiency. Numerical experiments show that NGD avoids local minima in both Born machines and in general MPS tomography. Remarkably, we show that NGD with line search can converge to the global minimum in only a few iterations. Second, for the BM algorithm, we introduce a warm-start initialization based on the TTNS-Sketch algorithm. We show that gradient descent under a warm initialization does not encounter the causality trap and admits rapid convergence to the ground truth.

Figures

Figures reproduced from arXiv: 2505.06419 by the authors.

Figure 2
Figure 2. FIG. 2. Graphical representations of the underlying model [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Performance of Born machine algorithm for the periodic spin system model in Equation (4). The models are initialized [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. FIG. 4. Plot of total variation (TV) distance of the trained Born machine model with respect to the true model [PITH_FULL_IMAGE:figures/full_fig_p003_4.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: FIG. 5. Performance of MPS tomography algorithm for the [PITH_FULL_IMAGE:figures/full_fig_p004_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Performance of NGD for MPS tomography algorithm [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 8
Figure 8. Figure 8: FIG. 8. Performance of NGD for the Born machine algorithm on the periodic spin system model in Equation (4). The models [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]
Figure 9
Figure 9. Figure 9: FIG. 9. Performance of Born machine algorithm for the periodic spin system model. The models are initialized with TTNS [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Variational inference and density estimation with non-negative tensor train

    math.NA 2025-07 conditional novelty 6.0 of 10

    A two-stage algorithm that fits non-negative tensor trains to high-dimensional discrete distributions using Newton-based alternating minimization with a log barrier converges much faster than the previous multiplicati...

Reference graph

Works this paper leans on

38 extracted references · 24 canonical work pages · cited by 1 Pith paper

  1. [1]

    Larochelle and I

    H. Larochelle and I. Murray, inProceedings of the four- teenth international conference on artificial intelligence and statistics(JMLR Workshop and Conference Pro- ceedings, 2011) pp. 29–37

    2011

  2. [2]

    Germain, K

    M. Germain, K. Gregor, I. Murray, and H. Larochelle, in International conference on machine learning(PMLR,

  3. [3]

    Salimans, A

    T. Salimans, A. Karpathy, X. Chen, and D. P. Kingma, arXiv preprint arXiv:1701.05517 (2017)

    arXiv 2017

  4. [4]

    D. P. Kingma and M. Welling, arXiv preprint arXiv:1312.6114 (2013)

    arXiv 2013

  5. [5]

    LeCun, S

    Y. LeCun, S. Chopra, R. Hadsell, M. Ranzato, and F. Huang, Predicting structured data1(2006)

    work page 2006

  6. [6]

    Rezende and S

    D. Rezende and S. Mohamed, inInternational conference on machine learning(PMLR, 2015) pp. 1530–1538

    2015

  7. [7]

    Goodfellow, J

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Ben- gio, Advances in neural information processing systems 27(2014)

    work page 2014

  8. [8]

    Y. Song, J. Sohl-Dickstein, D. P. Kingma, A. Kumar, S. Ermon, and B. Poole, arXiv preprint arXiv:2011.13456 (2020)

    arXiv 2020

Show all 38 references
  1. [9]

    Z.-Y. Han, J. Wang, H. Fan, L. Wang, and P. Zhang, Physical Review X8, 031012 (2018)

    2018

  2. [10]

    Y. Hur, J. G. Hoskins, M. Lindsey, E. M. Stoudenmire, and Y. Khoo, arXiv preprint arXiv:2202.11788 (2022)

    2022 arXiv

  3. [11]

    X. Tang, Y. Hur, Y. Khoo, and L. Ying, Research in the Mathematical Sciences10, 19 (2023)

    2023

  4. [12]

    G. S. Novikov, M. E. Panov, and I. V. Oseledets, in Uncertainty in Artificial Intelligence(PMLR, 2021) pp. 1321–1331

    2021

  5. [13]

    S. R. White, Physical review letters69, 2863 (1992)

    1992

  6. [14]

    Glasser, R

    I. Glasser, R. Sweke, N. Pancotti, J. Eisert, and I. Cirac, Advances in neural information processing systems32 (2019)

    2019

  7. [15]

    Cramer, M

    M. Cramer, M. B. Plenio, S. T. Flammia, R. Somma, D. Gross, S. D. Bartlett, O. Landon-Cardinal, D. Poulin, and Y.-K. Liu, Nature communications1, 149 (2010)

    2010

  8. [16]

    Lanyon, C

    B. Lanyon, C. Maier, M. Holz¨ apfel, T. Baumgratz, C. Hempel, P. Jurcevic, I. Dhand, A. Buyskikh, A. Daley, M. Cramer,et al., Nature Physics13, 1158 (2017)

    2017

  9. [17]

    Mitarai, M

    K. Mitarai, M. Negoro, M. Kitagawa, and K. Fujii, Phys- ical Review A98, 032309 (2018). 9

    2018

  10. [18]

    Bergholm, J

    V. Bergholm, J. Izaac, M. Schuld, C. Gogolin, S. Ahmed, V. Ajith, M. S. Alam, G. Alonso-Linaje, B. AkashNarayanan, A. Asadi,et al., arXiv preprint arXiv:1811.04968 (2018)

    2018 arXiv

  11. [19]

    Schuld, V

    M. Schuld, V. Bergholm, C. Gogolin, J. Izaac, and N. Kil- loran, Physical Review A99, 032331 (2019)

    2019

  12. [20]

    Glorot and Y

    X. Glorot and Y. Bengio, inProceedings of the thir- teenth international conference on artificial intelligence and statistics(JMLR Workshop and Conference Pro- ceedings, 2010) pp. 249–256

    2010

  13. [21]

    Pascanu, T

    R. Pascanu, T. Mikolov, and Y. Bengio, inInternational conference on machine learning(PMLR, 2013) pp. 1310– 1318

    2013

  14. [22]

    Erhan, A

    D. Erhan, A. Courville, Y. Bengio, and P. Vincent, in Proceedings of the thirteenth international conference on artificial intelligence and statistics(JMLR Workshop and Conference Proceedings, 2010) pp. 201–208

    2010

  15. [23]

    J. R. McClean, S. Boixo, V. N. Smelyanskiy, R. Babbush, and H. Neven, Nature communications9, 1 (2018)

    2018

  16. [24]

    C. O. Marrero, M. Kieferov´ a, and N. Wiebe, PRX Quan- tum2, 040316 (2021)

    2021

  17. [25]

    Cerezo, A

    M. Cerezo, A. Sone, T. Volkoff, L. Cincio, and P. J. Coles, Nature communications12, 1 (2021)

    2021

  18. [26]

    Cerezo and P

    M. Cerezo and P. J. Coles, Quantum Science and Tech- nology6, 035006 (2021)

    2021

  19. [27]

    E. R. Anschuetz and B. T. Kiani, arXiv preprint arXiv:2205.05786 (2022)

    2022 arXiv

  20. [28]

    E. R. Anschuetz, inInternational Conference on Learn- ing Representations(2021)

    2021

  21. [29]

    Dborin, F

    J. Dborin, F. Barratt, V. Wimalaweera, L. Wright, and A. G. Green, Quantum Science and Technology7, 035014 (2022)

    2022

  22. [30]

    Vogel and H

    K. Vogel and H. Risken, Physical Review A40, 2847 (1989)

    1989

  23. [31]

    Leonhardt, Physical review letters74, 4101 (1995)

    U. Leonhardt, Physical review letters74, 4101 (1995)

    1995

  24. [32]

    A. G. White, D. F. James, P. H. Eberhard, and P. G. Kwiat, Physical review letters83, 3103 (1999)

    1999

  25. [33]

    Torlai, G

    G. Torlai, G. Mazzola, J. Carrasquilla, M. Troyer, R. Melko, and G. Carleo, Nature physics14, 447 (2018)

    2018

  26. [34]

    Haegeman, J

    J. Haegeman, J. I. Cirac, T. J. Osborne, I. Piˇ zorn, H. Ver- schelde, and F. Verstraete, Physical review letters107, 070601 (2011)

    2011

  27. [35]

    Haegeman, C

    J. Haegeman, C. Lubich, I. Oseledets, B. Vandereycken, and F. Verstraete, Physical Review B94, 165116 (2016)

    2016

  28. [36]

    Oseledets and E

    I. Oseledets and E. Tyrtyshnikov, Linear Algebra and its Applications432, 70 (2010). Appendix A: Characterization of the causality trap We shall show that the causality trap can be charac- terized as an inability of a trained MPS model to capture boundary correlations. We cons...

    2010

  29. [37]

    Therefore,D z(pcausal) is the only term contributing to the NLL gap. Moreover, sinceX 1,Xn are close to being independent inp causal, it follows thatp causal (x1,xn) ≈ pcausal x1 pcausal xn =p ⋆ x1p⋆ xn, where the second equality follows from the symmetry ofp ⋆ andp causal. Th...

  30. [38]

    One likely explanation for the 2-site DMRG method encountering 12 local minima issues is that the truncation step in the 2-site DMRG is not variational

    by iterating fromi= 1 toi=n−1. One likely explanation for the 2-site DMRG method encountering 12 local minima issues is that the truncation step in the 2-site DMRG is not variational. The factorization Equa- tion (D4) does not necessarily minimize the lossL, and it instead sim...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.